Financial Outcomes for Debt Settlement Programsú

Will S. Dobbie Harvard Kennedy School

May 2, 2020

1. Summary

Debt settlement programs are designed to facilitate the less-than-full-balance resolution of portfolios of unsecured debt owed by individuals who are experiencing significant financial diculty and who lack the resources to satisfy, in full, their current debt obligations. This report examines the financial outcomes of individuals enrolling in a debt settlement program using detailed data for a sample population of approximately 110,000 individuals and over 735,000 accounts. We also examine the methodology, analyses, and results described in a series of reports by Hemming Morse, LLP (Greg Regan) that also examine the financial outcomes of individuals enrolling in a debt settlement program to determine whether the data support the Regan reports’ analyses and conclusions. Our analysis reveals that individuals starting a debt settlement program have, on average, over $30,000 of unsecured debt across 6.68 accounts. Seventy-six percent of these individuals successfully settle at least one account through the debt settlement program over the first 36 months, with these individuals settling an average of 3.37 accounts and approximately 51 percent of their enrolled debt through the program during this time. Settled accounts yield an average savings of about $1,700 based on the current balance and after accounting for fees, with individuals saving an average of just under $5,800 on settled accounts based on the current balance and after fees. We can also express the debt reduction in terms of the fraction of the current balance that is forgiven, which more easily facilitates comparisons with other forms of debt relief. These calculations show that individuals enrolling in debt settlement programs receive an average debt write-down of 33.2 percent on settled accounts after accounting for fees. All of our findings are substantively identical to those in the most recent iteration of the report by Hemming Morse, LLP (Greg Regan). Reviewing the evidence from both this report and recent research, we conclude that debt settlement programs have the potential to significantly benefit many financially distressed individuals, particularly if they are not eligible for Chapter 7 bankruptcy protection or wish to avoid the negative consequences of Chapter 13 bankruptcy restructuring.

úThis report was commissioned by the American Fair Credit Council (AFCC), the trade association representing the debt settlement industry. The AFCC reviewed this report for accuracy and to ensure that no personally identifiable information was disclosed. The report’s methodology, analysis, and results were not moderated or mediated by the AFCC. The opinions, findings, and conclusions expressed in this report are those of the author and not those of the AFCC or any other organization. Information on the author and his qualifications is available at https://www.hks. harvard.edu/faculty/will-dobbie.

1 2. Background and Data

2.1. The Debt Settlement Industry

Debt settlement firms are for-profit organizations that negotiate with unsecured creditors on behalf of their clients. Firms typically negotiate a full discharge of an individual’s unsecured debt in exchange for either a one-time lump-sum payment or a series of smaller payments. These settlements generally represent a partial write-oby the creditor of the current balance at the time of settlement. Debt settlement firms do not provide financial counseling, legal, tax or bankruptcy advice, or help with secured debts such as mortgages or auto loans. The debt settlement process typically begins with a comprehensive phone screen. To be eligible for a debt settlement program, firms generally require that individuals have at least $10,000 to $15,000 in unsecured debt, a steady source of income, no recent bankruptcy filings, and live in a state where the firm operates. Following the phone screen, individuals interested in starting a debt settlement program are sent a request for financial documentation along with a contract that details the firm’s services and fees, and that contains all of the required regulatory disclosures. Individuals who return the required documentation, submit the contract, and pass the firm’s enrollment criteria are then enrolled in a debt settlement program and begin making monthly or semi-monthly deposits into a dedicated account, a separate bank account owned by the individual and not accessible by the debt settlement firm.1 Negotiations with creditors usually begin after sucient funds have accumulated in the client’s dedicated account, generally around 20 percent of any given debt. Debt settlement firms collect fees on a per-debt basis directly from clients. By law, three conditions must be met before any per-debt fees can be collected from clients: (1) the firm must successfully negotiate the terms of settlement for a given debt, (2) the client must agree to the terms of the negotiated settlement for the debt, and (3) the client must make at least one payment to the relevant creditor as a result of the negotiated settlement agreement.

2.2. The Regan Report

The best available information on individuals enrolling in a debt settlement program comes from a series of reports by Hemming Morse, LLP (authored by Greg Regan, a partner in the ﬁrm), that have been commissioned by the AFCC. The most recent version of the “Regan report,” completed in 2017, includes information on individuals enrolling in a debt settlement program between January 1, 2011 and March 31, 2017. The 2017 report includes detailed breakdowns of the post-enrollment outcomes of these individuals, as well as a comparison of these post-enrollment outcomes with the hypothetical outcomes that would have occurred if these individuals had only made the minimum payment on their unsecured debt and not enrolled in a debt settlement program. We were commissioned by the AFCC to conduct an independent examination of the 2017 iteration of the Regan report to determine whether the available data supported the report’s analyses and conclusions. We were provided with the account-level data used in the Regan report, as well as

1Debt settlement deposits are used only for payments to creditors and the debt settlement ﬁrm’s fees. The payment processors who administer these dedicated accounts commonly charge $5-$10 per month to maintain the account over the program life (usually up to 48 months). We focus exclusively on the much larger per-debt fees when calculating program outcomes below.

2 detailed notes on the sample restrictions and analyses used in the report and full access to the author to discuss the report’s methodology, scope, and outcomes. Per the disclaimer on the ﬁrst page, the AFCC reviewed this report for accuracy and to ensure that no personally identiﬁable information was disclosed but did not moderate or mediate our methodology, analysis, or results.

2.3. Data

The debt settlement records used in our analysis come from the AFCC. The data contain the universe of accounts enrolled in debt settlement programs between January 1, 2011 and March 31, 2017 for five of the largest debt settlement firms in the industry. We observe information on the starting balance for each account, the current balance for each account, the settlement amount for each settled account, and any charged fees for each settled account. We focus on the per-debt fees throughout this analysis, ignoring the $5-$10 monthly cost of maintaining the dedicated debt settlement account. We use these data to calculate account- and individual-level statistics for outcomes after individuals enter a debt settlement program over a range of time horizons. Shorter time horizons allow us to include more individuals in our sample population, while longer time horizons allow us to examine outcomes that occur further in the future. Our preferred sample population examines outcomes in the first 36 months after individuals enter a debt settlement program, ignoring outcomes that occur after the first 36 months even though there may be additional settlements in subsequent months (particularly for larger debts). This 36-month window balances the need for a larger sample with the need for a longer time horizon. Appendix Table 1 and Appendix Figures 1 and 2 provide results at both shorter and longer time horizons and are discussed throughout. At the account level, we calculate measures of the debt reduction for each settled account before accounting for per-debt fees, the debt reduction for each settled account after accounting for per-debt fees, and the debt reduction divided by the per-debt fee amount for each settled account. We also aggregate these variables across all of an individual’s accounts, yielding individual-level measures of the number of settled accounts, the total debt reduction for each settled account before accounting for per-debt fees, the total debt reduction for each settled account after accounting for per-debt fees, and the average debt reduction divided by the per-debt fee amount for each settled account. All individual- level statistics are weighted by the starting balance in each account, such that larger accounts matter more for these individual-level averages. We impose three sample restrictions to arrive at our preferred sample population. First, we drop 353,662 individuals (2,696,275 accounts) enrolling in a debt settlement program after March 31, 2014 to ensure that we observe all individuals for at least 36 months after entering a debt settlement program. Second, we drop 10,837 individuals (70,219 accounts) that failed to make their first draft deposit, as these individuals never functionally enrolled in a debt settlement program. Finally, we drop 3,464 individuals (25,683 accounts) where at least one account has missing or conflicting information. None of our results substantively change if we relax these sample restrictions. The final dataset for our preferred sample population includes 110,157 individuals and 735,382 accounts.

3 3. Results

3.1. Descriptive Statistics

Table 1 provides descriptive statistics for our preferred sample population at both the account and individual levels. Panel A shows that individuals enrolling in a debt settlement program are in considerable financial distress at the beginning of the program. Individuals enroll an average of 6.68 accounts, with the typical account balance equaling $4,495. Individuals therefore enroll an average of $30,008 of unsecured debt across all accounts ( 6.68 $4,495). We find nearly identical results ¥ ◊ for sample populations where we measure outcomes over both shorter and longer time horizons, as reported in Appendix Table 1. By comparison, the typical not-for-profit debt management plan includes about $20,000 in unsecured debt (Dobbie and Song, 2019), while the typical Chapter 7 or Chapter 13 consumer bankruptcy filing includes more than $90,000 in unsecured debt (Auclert, Dobbie, and Goldsmith-Pinkham, 2019). Panel B shows that 50 percent of enrolled accounts are settled during our 36-month sample window, with 76 percent of individuals settling at least one account during this period. Individuals settle an average of 3.37 accounts and approximately 51 percent of their enrolled debt during this period. Fifty- four percent of individuals settle over 50 percent of their enrolled debt during this period, 37 percent settle over 75 percent of their enrolled debt, and 18 percent settle all of their enrolled debt. We find similar results for sample populations where we measure outcomes over both shorter and longer time horizons, as reported in Appendix Table 1. Panel C further shows that these settlements are reached, on average, about 13.8 months after the start of the debt settlement program, with the first settlement usually occurring four to five months after the start of the program. Starting account balances on these settled accounts average $4,667, just slightly more than the average starting balance on all accounts. Current balances at the time of the settlement average $5,174, meaning the average settled account increases by $507, or 10.9 percent, from enrollment to settlement due to interest and late fees. The average settlement amount is $2,437, substantially lower than both the average starting and current balances ($4,667 and $5,174, respectively). Gross fees average $1,018 for these settled accounts, or approximately 21.8 percent of the enrolled balance and 19.7 percent of the balance at the time of settlement. Individuals therefore have an average starting balance of $15,731 on settled accounts ( 3.37 $4,667), an average current ¥ ◊ balance of $17,440 ( 3.37 $5,174), an average settlement amount of $8,215 ( 3.37 $2,437), and ¥ ◊ ¥ ◊ average fees of $3,432 ( 3.37 $1,018). We again find similar results for sample populations where ¥ ◊ we measure outcomes over both shorter and longer time horizons, as reported in Appendix Table 1. Putting these results together, Panel D shows that the typical settlement during our 36-month sample window results in $2,737 in debt reduction before accounting for fees, and $1,718 in net savings after accounting for fees. Individuals therefore see an average debt reduction of $9,224 on settled accounts before accounting for fees ( 3.37 $2,737), and $5,792 in net savings after accounting for ¥ ◊ fees ( 3.37 $1,718). Settled accounts yield an average of $2.76 in debt reduction per dollar of ¥ ◊ fees, with the individuals receiving an average of $2.67 in debt reduction per dollar of fees on settled accounts. We, once again, find similar results for sample populations where we measure outcomes over both shorter and longer time horizons, as reported in Appendix Table 1. We can also express the debt reduction in terms of the fraction of the current balance that is

4 forgiven, which more easily facilitates comparisons with other forms of debt relief. These calculations show that individuals enrolling in debt settlement programs receive an average debt write-down of 33.2 percent on settled accounts after accounting for fees. By comparison, individuals enrolling in not-for- profit debt management plans generally do not receive a write-down on the original principal, though interest payments and late fees are significantly reduced (Dobbie and Song, 2019). Individuals filing for bankruptcy protection receive an average debt-write down of more than 90 percent for unsecured debt included in a Chapter 7 discharge and 60 to 70 percent for unsecured debt included in a Chapter 13 discharge (Dobbie, Goldsmith-Pinkham, and Yang, 2017). All of the statistics reported here are substantively identical to those reported in the 2017 Regan report.

3.2. Relationship with Baseline Characteristics

Figures 1 and 2 present estimates of the relationship between post-enrollment financial outcomes and baseline characteristics measured at the time of program enrollment. To construct each plot, we first split each baseline characteristic into 20 equally-sized bins. We then create indicator variables for each of these bins, omitting the smallest value. We finally regress each post-enrollment outcome on these indicator variables, along with the other baseline controls, and plot the results. These plots therefore represent the non-parametric relationship between each post-enrollment outcome and the baseline characteristics. Following Table 1, we measure all outcomes for the first 36 months after individuals enter a debt settlement program and report results over both shorter and longer time horizons in Appendix Figures 1 and 2. We find that post-enrollment outcomes are generally more positive for individuals with higher levels of debt and, perhaps, for individuals with more accounts. Figure 1 shows that the probability that an account is settled in the first 36 months increases sharply with the starting account balance amount until about $5,000, leveling oafter this threshold. The probability that an account is settled also generally increases with the total number of enrolled accounts. Debt reduction on settled accounts, both before and after accounting for fees, is mechanically increasing with the starting balance but unrelated to the total number of enrolled accounts. Figure 2 shows that there is a similar pattern of results at the individual level, with the probability that any account is settled increasing with the total starting balance amount until about $40,000 and leveling oafter this threshold. Debt reduction on settled accounts, both before and after accounting for fees, is again mechanically increasing with the starting balance but unrelated to the total number of enrolled accounts. We again find nearly identical results when measuring outcomes over both shorter and longer time horizons, as reported in Appendix Figures 1 and 2.

3.3. Accretion of Accounts in Debt Settlement Programs

The 2017 Regan report also calculates the amount of debt that individuals would have accumulated had they only made the minimum payment on their credit cards and not enrolled in a debt settlement program. We successfully replicated these calculations, which we only brieﬂy describe here for completeness. The annualized accretion rate for enrolled debt falls from more than 25 percent in the ﬁrst few months after program enrollment to less than 10 percent after one and a half years. By comparison, the annualized accretion rate for credit card debt remains at or above 15 percent after

5 one and a half years when making the minimum required payment on a credit card. Individuals are therefore expected to have lower debt levels, even accounting for accretion, after enrolling in a debt settlement program compared to making the minimum required payment on a credit card. One potential critique of these calculations is that they implicitly assume individuals cannot aord the minimum required payment on their credit cards. To the extent that individuals can repay significantly more than the minimum required payment, the benefit of a debt settlement program is substantially reduced or eliminated, as noted by a recent report by the Center for Responsible Lending (Coleman and Stifler, 2015). But this critique ignores the fact that financially distressed individuals are often caught in a debt trap where their outcomes are already deteriorating. Dobbie and Song (2015) and Dobbie, Goldsmith-Pinkham, and Yang (2017) document, for example, that individuals filing for Chapter 13 bankruptcy protection start to see a deterioration of their outcomes about one to two years before filing. Outcomes continue to deteriorate, often sharply, for the roughly half of individuals whose Chapter 13 filings are dismissed by the court. We therefore believe that the assumption that at least some financially distressed individuals cannot aord the minimum required payment on their credit cards is reasonable, as are the conclusions based on this assumption.

4. Conclusion

The results from our analysis show that 76 percent of individuals successfully settle at least one account in the first 36 months after enrolling in a debt settlement program, with settled accounts yielding an average debt write-down of 33.2 percent after accounting for fees. By comparison, individuals enrolling in not-for-profit debt management plans generally do not receive a write-down on the original principal and individuals filing for bankruptcy protection receive an average debt-write down of more than 90 percent for unsecured debt included in a Chapter 7 discharge and about 60 to 70 percent for unsecured debt included in a Chapter 13 discharge. Based on these findings, we can reasonably conclude that debt settlement programs have the potential to benefit many financially distressed individuals, particularly if they are not eligible for or interested in Chapter 7 or Chapter 13 bankruptcy protection. An important limitation of our analysis is that we only observe outcomes for individuals starting a debt settlement program. Our estimates based on these data may not represent the causal impact of starting a debt settlement program, as the outcomes for these individuals may have changed over time even if they had not started a debt settlement program. Obtaining the causal impact of starting a debt settlement program would likely require comparing the outcomes of individuals who start a debt settlement program to the outcomes of otherwise similarly situated individuals who did not start a debt settlement program, as would be the case in a randomized control trial. Estimating such a causal eect is an important area for future work as it will allow us to improve on the hypothetical calculations in the most recent iteration of the Regan report.

6 References

[1] Auclert, Adrien, Will Dobbie, and Paul Goldsmith-Pinkham. 2019. “Macroeconomic Eects of Debt Relief: Consumer Bankruptcy Protections in the Great Recession.” NBER Working Paper No. 25685.

[2] Coleman, Brandon, and Lisa Stiﬂer. 2015. “CRL Finds AFCC Debt Settlement Report Falls Short.” Center for Responsible Lending Issue Brief.

[3] Dobbie, Will, Paul Goldsmith-Pinkham, and Crystal Yang. 2017. “Consumer Bankruptcy and Financial Health.” Review of Economics and Statistics.

[4] Dobbie, Will, and Jae Song. 2015. “Debt Relief and Debtor Outcomes: Measuring the Eects of Consumer Bankruptcy Protection.” American Economic Review.

[5] Dobbie, Will, and Jae Song. 2019. “Targeted Debt Relief and the Origins of Financial Distress: Experimental Evidence from Distressed Credit Card Borrowers.” American Economic Review.

[6] Regan, Greg. 2017. “Options for Consumers in Crisis: An Updated Economic Analysis of The Debt Settlement Industry.” American Fair Credit Council Report.

7 Table 1: Descriptive Statistics

Accounts Individual Statistics Statistics Panel A: Enrollment Variables (1) (2) Number of Accounts – 6.68 Starting Balance 4,495.19 30,008.80

Panel B: Outcome Variables for All Accounts Any Account Settled 0.50 0.76 Number of Accounts Settled – 3.37 Fraction of Accounts Settled – 0.51 50 Percent of Accounts Settled – 0.54 75 Percent of Accounts Settled – 0.37 100 Percent of Accounts Settled – 0.18

Panel C: Outcomes for Settled Accounts Avg. Months to Settlement 13.81 12.04 Starting Balance 4,667.80 15,731.50 Current Balance 5,174.93 17,440.65 Settlement Amount 2,437.79 8,215.88 Gross Fees 1,018.36 3,432.10

Panel D: Normalized Outcomes for Settled Accounts Debt Reduction 2,737.14 9,224.77 Debt Reduction Less Fees 1,718.78 5,792.66 Debt Reduction Over Fees 2.76 2.67 Observations 735,382 110,157 Notes. This table reports descriptive statistics at the account and individual levels. The sample includes both accounts and individuals enrolled in a debt settlement program between January 1, 2011 and March 31, 2014 meeting the sample criteria described in the text. Baseline characteristics are measured at program enrollment and outcomes are measured for the ﬁrst 36 months after program enrollment. Column 1 reports accounts statistics. Column 2 reports individual statistics weighted by the starting balance in each account. Debt reduction is the current balance less the settlement amount. Debt reduction less fees is the current balance less the settlement amount less the gross fees. Debt reduction over fees is the current balance less the settlement amount divided by the gross fees.

8 Figure 1: Account Outcomes and Baseline Characteristics Account Settled & Account Settled & Starting Balance Number of Accounts .6 .6 .5 .5 Account Settled Account Settled .4 .4 0 5000 10000 15000 20000 0 5 10 15 20 Starting Account Balance Number of Accounts

Debt Reduction & Debt Reduction & Starting Balance Number of Accounts 15000 15000 10000 10000 Debt Reduction Debt Reduction 5000 5000 0 0

0 5000 10000 15000 20000 0 5 10 15 20 Starting Account Balance Number of Accounts

Debt Reduction Less Fees & Debt Reduction Less Fees & Starting Balance Number of Accounts 8000 8000 6000 6000 4000 4000 Debt Reduction Less Fees Debt Reduction Less Fees 2000 2000 0 0

0 5000 10000 15000 20000 0 5 10 15 20 Starting Account Balance Number of Accounts

Notes. This figure reports the relationship between post-enrollment account outcomes and baseline characteristics. The sample includes both accounts and individuals enrolled in a debt settlement program between January 1, 2011 and March 31, 2014 meeting the sample criteria described in the text. Baseline characteristics are measured at program enrollment and outcomes are measured for the first 36 months after program enrollment. We construct each plot by first splitting each characteristic into 20 equally-sized bins. We then create indicator variables for each of these bins, omitting the smallest value. We finally regress each post-enrollment outcome on these indicator variables. The blue line in each figure represents the relationship between the listed account outcome and characteristic with no additional controls. The red line in each figure represents the relationship between the listed account outcome and characteristic controlling for the other listed characteristics and monthly income. Account settled is an indicator for the account being settled. Debt reduction is the current balance less the settlement amount. Debt reduction less fees is the current balance less the settlement amount less the gross fees.

9 Figure 2: Individual Outcomes and Baseline Characteristics Any Account Settled & Any Account Settled & Starting Client Balance Number of Accounts .9 .9 .8 .8 Account Settled Account Settled .7 .7 .6 .6 0 20000 40000 60000 80000 100000 0 5 10 15 20 Starting Client Balance Number of Accounts

Debt Reduction & Debt Reduction & Starting Client Balance Number of Accounts 30000 30000 20000 20000 Debt Reduction Debt Reduction 10000 10000 0 0

0 20000 40000 60000 80000 100000 0 5 10 15 20 Starting Client Balance Number of Accounts

Debt Reduction Less Fees & Debt Reduction Less Fees & Starting Client Balance Number of Accounts 20000 20000 15000 15000 10000 10000 Debt Reduction Less Fees Debt Reduction Less Fees 5000 5000 0 0

0 20000 40000 60000 80000 100000 0 5 10 15 20 Starting Client Balance Number of Accounts

Notes. This figure reports the relationship between post-enrollment individual outcomes and baseline characteristics. The sample includes both accounts and individuals enrolled in a debt settlement program between January 1, 2011 and March 31, 2014 meeting the sample criteria described in the text. Baseline characteristics are measured at program enrollment and outcomes are measured for the first 36 months after program enrollment. We construct each plot by first splitting each characteristic into 20 equally-sized bins. We then create indicator variables for each of these bins, omitting the smallest value. We finally regress each post-enrollment outcome on these indicator variables. The blue line in each figure represents the relationship between the listed program outcome and characteristic with no additional controls. The red line in each figure represents the relationship between the listed program outcome and characteristic controlling for the other listed characteristics and monthly income. Account settled is an indicator for any account being settled. Debt reduction is the current balance less the settlement amount. Debt reduction less fees is the current balance less the settlement less the gross fees.

10 Appendix Table 1: Descriptive Statistics at Dierent Time Horizons

Accounts Statistics Client Statistics 24 36 48 60 24 36 48 60 Months Months Months Months Months Months Months Months Panel A: Enrollment Variables (1) (2) (3) (4) (5) (6) (7) (8) Number of Accounts – – – – 6.90 6.68 6.43 6.32 Starting Balance 4,364.63 4,495.19 4,665.38 4,900.04 30,132.38 30,005.86 30,005.86 30,948.42

Panel B: Outcome Variables for All Accounts Any Account Settled 0.45 0.50 0.52 0.51 0.76 0.75 0.75 0.75 Number of Accounts Settled – – – – 3.13 3.37 3.32 3.24 Fraction of Accounts Settled – – – – 0.46 0.51 0.52 0.51 50 Percent of Accounts Settled – – – – 0.49 0.54 0.53 0.52 75 Percent of Accounts Settled – – – – 0.26 0.37 0.39 0.38 100 Percent of Accounts Settled – – – – 0.13 0.18 0.22 0.22

Panel C: Outcomes for Settled Accounts

11 Avg. Months to Settlement 12.85 13.81 14.43 14.42 11.41 12.39 12.39 12.35 Starting Balance 4,578.95 4,667.80 4,798.41 4,948.72 14,336.03 15,943.06 15,943.06 16,042.79 Current Balance 5,069.06 5,174.93 5,318.76 5,487.10 15,870.48 17,671.94 17,671.94 17,788.13 Settlement Amount 2,414.31 2,437.79 2,513.27 2,606.58 7,558.85 8,350.50 8,350.50 8,450.04 Gross Fees 1,002.38 1,018.36 1,038.78 1,049.99 3,138.30 3,451.42 3,451.42 3,403.87

Panel D: Normalized Outcomes for Settled Accounts Debt Reduction 2,654.75 2,737.15 2,805.49 2,880.52 8,311.63 9,321.44 9,321.44 9,338.09 Debt Reduction Less Fees 1,652.37 1,718.78 1,766.71 1,830.53 5,173.33 5,870.01 5,870.01 5,934.23 Debt Reduction Over Fees 2.72 2.76 2.79 2.89 2.63 2.69 2.69 2.78 Observations 1,149,408 735,382 407,905 152,644 166,490 110,157 63,422 24,168 Notes. This table reports descriptive statistics at dierent time horizons. The sample includes both accounts and individuals enrolled in a debt settlement program between January 1, 2011 and March 31, 2017 over the indicated time horizon. Columns 1-4 reports accounts statistics. Columns 5-8 reports client statistics weighted by the starting balance in each account. Debt reduction is the current balance minus the settlement amount. Debt reduction less fees is the current balance minus the settlement amount minus the gross fees. Debt reduction over fees is the current balance minus the settlement amount divided by the gross fees. Appendix Figure 1: Account Outcomes and Baseline Characteristics at Dierent Time Horizons Account Settled & Account Settled & Starting Balance Number of Accounts .6 .6 .4 .4 Account Settled Account Settled .2 .2 0 5000 10000 15000 20000 0 5 10 15 20 Starting Account Balance Number of Accounts